大学生干眼风险列线图预测模型的构建及验证

Construction and Verification of Dry Eye Risk Nomogram Model for College Students

目的:分析大学生近视分级与干眼程度的相关性,探索列线图模型对大学生干眼患病风险的预测价值。方法:横断面研究。于2023年2月随机收集武汉大学就读的365例大学生为研究对象,通过问卷调查、眼表综合分析仪、角膜荧光素染色检查采集相关数据。运用Logistic回归分析筛选大学生干眼相关因素,构建大学生干眼发病风险列线图预测模型并进行内部验证。结果:共纳入365名受检者,年龄(22.8±2.9)岁,男性占47.7%,干眼的患病率为40.0%(146/365)。多因素Logistic回归分析显示,吸烟(OR=2.45,P=0.045)、视频终端使用时长(OR=4.09,P<0.001)、睡眠时长(OR=2.02,P=0.004)、抑郁(中度vs.无,OR=3.67,P=0.001;重度vs.无,OR=4.66,P=0.030)、过敏性结膜炎(OR=2.38,P=0.036)、屈光不正(<-3.0 D vs.>-0.5 D,OR=4.45,P<0.001)、配戴角膜接触镜(OR=4.84,P<0.001)为大学生干眼的相关因素,性别、年龄、身份、婚姻、体质量指数、饮酒、运动、口服避孕药、干眼家族史不是大学生干眼的相关因素(P>0.05)。屈光度数与眼表疾病指数问卷(r=-0.73,P<0.001)及泪膜破裂时间(r=-0.64,P<0.001)呈高度相关,与泪河高度(r=-0.47,P<0.001)呈中度相关,与睑板腺丢失(r=-0.24,P<0.001)及堵塞程度(r=-0.20,P<0.001)相关性较低。由7项危险因素构建的干眼列线图预测模型,Hosmer-Lemeshow检验得P=0.150,二元Logistic模型与观测值拟合良好,C指数为0.794(95%置信区间:0.746~0.841),受试者工作特征曲线下面积为0.794(95%置信区间:0.745~0.842,P<0.001),敏感性78.1%,特异性74.9%,该模型具有中等预测能力。结论:中高度近视与大学生干眼显著相关,吸烟、视频终端使用时长、睡眠时长、抑郁、过敏性结膜炎、屈光不正、配戴角膜接触镜为大学生干眼的相关因素,基于此构建的风险预测模型可辅助早期筛查及干预大学生干眼。

Objective:To analyze the correlation between myopia and dry eye severity in college students,and construct a dry eye risk nomogram model in college students.Methods:This was a cross-sectional study.College students who were enrolled at Wuhan University in February 2023 were randomly selected as subjects,and conducted a questionnaire survey,keratograph 5M,and corneal fluorescence staining.Logistic regression analysis was employed to identify factors related to dry eye among college students,and a nomogram model for dry eye among college students was constructed and validated internally.Results:A total of 365 subjects were included,aged 22.8±2.9 years old,with 47.7%of males.The incidence of dry eye was 40.0%(146/365).Multivariate Logistic regression analysis showed that smoking(OR=2.45,P=0.045),video display terminal(VDT)(OR=4.09,P<0.001),sleep(OR=2.02,P=0.004),depression(moderate compared to no,OR=3.67,P=0.001;severe compared to no,OR=4.66,P=0.030),allergic conjunctivitis(OR=2.38,P=0.036),refractive error(<-3.0 D compared to>-0.5 D,OR=4.45,P<0.001),and wearing contact lenses(OR=4.84,P<0.001)were related to dry eye of college students.Gender,age,identity,marriage,boaly mass index,drinking,sports,oral contraceptives,and dry eye family history were not related to dry eye of college students(P>0.05).The refractive index was highly correlated with OSDI(r=-0.73,P<0.001)and NIBUT(r=-0.64,P<0.001),moderately correlated with TMH(r=-0.47,P<0.001),and less correlated with the degree of meibomian gland loss(r=-0.24,P<0.001)and blockage(r=-0.20,P<0.001).A dry eye nomogram prediction model was constructed from 7 risk factors.The P values of Hosmer Lemeshow test were 0.150,which is greater than 0.05,showing that this model fits well with the observed values.The C-index was 0.794(95%confidence interval:0.746-0.841),and the area under the ROC curve was 0.794(95%onfidence interval:0.745-0.842,P<0.001).The sensitivity was 78.1%,and the specificity was 74.9%.This model had moderate predictive ability.Conclusions:Moderate to high myopia is significantly associated with dry eye in college students.Smoking,VDT,sleep,depression,allergic conjunctivitis,refractive error,and wearing contact lenses are related factors for dry eye in college students.The nomogram model can assist in early screening and intervention of dry eye in college students.